1

Discursive Strategies of Legitimization: The Case of Abortion in Ireland in 2018

Jennifer O’Donovan, University of Edinburgh; Barbara Siller, University College Cork

Abstract

The following article is based on a study on abortion discourse carried out in Ireland in 2018 prior to and after the referendum to Repeal the Eighth Amendment to the Constitution, allowing for legislation to be introduced regulating termination of pregnancy. The main focus of the study is to identify the strategies of legitimization employed by both online users and campaign bodies in the period directly before and after the historic referendum. The article also aims to examine how the discursive strategies engage in creating national identities as well as identities of the collective voter groups. The corpus includes unregulated textual and regulated visual data collected between May and June 2018. The textual data derive from the Facebook pages of four prominent campaigning bodies encompassing both sides of the referendum one month before and after the referendum, while the visual data include photos taken of the campaign posters displayed in Ireland. Critical Discourse Analysis forms the conceptual framework of the study. Analysis was undertaken by employing Reyes’ (2011) model of strategies of legitimization in political discourse as well as the two visual analysis frameworks proposed by van Leeuwen (2008) and Ledin and Machin (2018). A mixed-method approach in the format of a triangulation design (Creswell & Plano Clark, 2007) is used, whereby qualitative textual and visual discourses are transformed into quantitative data for the purpose of analysis. The findings indicate that the strategy of emotion is the most utilized one; this becomes apparent through the quantitatively high use of terms such as ‘rape’, ‘incest’ or ‘murder’, operationalised to provoke an emotional reaction in the reader. The results also point to high instances of ‘othering’ strategies employed by the different groups participating in the discourse.

Keywords: Critical Discourse Analysis; abortion; identity; strategies of legitimization.

2

1. Introduction The Eighth Amendment to the Irish Constitution, passed by referendum in 1983, equivalates the right to life of the foetus to the right to life of the mother, thereby outlawing abortion. The opportunity to introduce this clause arose after a period of government upheaval had paved the way for the Pro-Life Amendment Campaign to assure there would be no immediate possibility to legislate for legal abortions.

There were three subsequent referenda in November 1992. The first sought to remove suicide as legal grounds for abortion after the ‘X Case’ (the X Case referred to the case of a 14-year- old girl, a victim of rape, who was refused the possibility to travel to England to procure an abortion) in which a Supreme Court ruling stated that a risk to a woman’s life could be a legitimate reason for a lawful termination of pregnancy. This referendum was not upheld. The Irish population was in favour of the following two referenda, the first being that the Eighth Amendment could not impede a woman’s right to travel to another country for an abortion, and the second being the right to access information about abortion. A further referendum in 2002 attempted to remove the risk of suicide as grounds for termination and this was, once again, defeated.1

With the 2018 referendum the people of Ireland could, for the first time, vote for or against the possibility to legislate for lawful abortion. This occurred in an age when social media was constantly gaining in popularity and influence thereby allowing people the possibility to express their opinions openly on a public forum. This has opened the door to debates, arguments and information-sharing.

While this study draws on previous studies, in particular on Smyth’s (2005) studies on the media coverage of the ‘X Case’ in 1992 in Ireland, in contrast to them, it includes data from social media, namely Facebook. The study thus focuses on both official discourse by campaign bodies, and less regulated online discourse. In addition to the strategies of legitimization, the study also analyses how the identities of the voter groups are constructed through the creation of a clear-cut ‘Yes’ group, in favour of legalising abortion and a ‘No’ group, against the legalisation. These identities are established linguistically through the practice of ‘othering’ by

1 For a comprehensive account of the timeline of the referendum and the abortion politics in Ireland from the early 1980s to date, see Field (2018). While Field places the referendum in the context of the marriage equality referendum of 2015, he argues that this referendum “deserves its own place in history” (p. 608) highlighting its significance. An activist perspective is offered by Carnegie and Roth (2019), both members of the ARC. Insights into the referendum outcome, drawing on theories of deliberative democracy, voting behaviour and generational change, are provided in the article by Elkink et al. (2020).

3 which a group constructs its own identity through the creation of an antithesis (Wodak, 2011, p. 62).

2. Theoretical Approach This study places itself in the field of Discourse Analysis, a method of inquiry that sees language as socially constructed and as embedded in a political and ideological communicative context. In his book Language and Power (1989), Fairclough highlights the following three aspects of language:

a) Language is an integral part of society and is not a separate entity; b) Language is a “social process”; c) Language is a “socially conditioned process”, that is, determined by other extralinguistic elements of society (p. 22).

He thus explicitly points to the fact that language is a social phenomenon and cannot be separated from any societal issue such as exclusionary processes, hierarchies and power. These characteristics, however, are not inherent to language, as Wodak and Meyer (2009) emphasise, when affirming that “power does not necessarily derive from language, but language can be used to challenge power, to subvert it, to alter distributions of power in the short and long term” (p. 10). According to Jäger (2001) discourse can be defined:

‘as the flow of knowledge - and/or all societal knowledge stored - throughout all time’..., which determines individual and collective doing and/or formative action that shapes society, thus exercising power. As such, discourses can be understood as material realities sui generis”. (p. 34)

Beside the importance of the duration and the context within which discourses are situated, the determinative and formative aspects of discourses are also highlighted, given their relevance for this study. Moreover, a broad definition of discourse is suggested, which also indicates the inclusion of non-verbal discourses.

Since this study investigates discourses which serve to “justify courses of actions” (Reyes, 2011, p. 781), it considers as productive the strategies of legitimization outlined by Reyes. The strategies employed to analyse the data are: legitimation through emotions, legitimation through a hypothetical future, legitimation through rationality, voices of expertise, and

4 altruism, whereby one further distinction is drawn between positive and negative voices of expertise as the data point towards this. Since the study also considers iconographic discourse, visual approaches as suggested by van Leeuwen, Ledin and Machin further inform the theoretical framework.

3. Methodology and Data Collection The study employs a mixed-method approach, transforming the qualitative textual and visual discourses into quantitative data for the purpose of analysis. This method draws on the ‘Triangulation Design: Data Transformation Model’ as outlined by Creswell and Plano Clark (2007, p. 65), whereby first, qualitative and quantitative data are collected and analysed. Subsequently, qualitative data are transformed into quantitative data in order to compare the two sets. Finally, the quantitative and the qualitative analysis is interpreted.

The language employed by online users and the discourses which underpinned the issue of abortion in 2018 in Ireland in the lead up to and the period after the referendum form one part of the data corpus of this research. Both the language used in online comments and the discourses employed in the official campaigns through posters are analysed. The second part of the data is the visual discourse in the posters.

The data consist of the following: twenty-eight photographs of the campaign posters and 754 comments collected (of which 494 were analysed) from the Facebook pages of organisations on both sides of the referendum campaign: ‘’, ‘Abortion Rights Campaign’, ‘Save the 8th’ and ‘the Life Institute’. The comments by readers responding to the posts were gathered in the month leading up to and the month after the referendum, i.e. in the period May- June 2018. The comments were analysed by transforming the qualitative data into quantitative data by assigning each comment to a strategy (or ‘other’ where none was applicable) with a numeric value of 1. This was done in order to identify the frequency of the strategies employed in constructing identities utilizing Reyes’ (2011) strategies of legitimization as a framework of reference. The frequency was then calculated as a percentage of the total number of comments for the voter group in the particular time period. Strategies of legitimation are intended as discursive strategies employed by social actors to “justify their political agenda to maintain or alter the direction of a whole nation” (Reyes, 2011, p. 783). The following five strategies are considered for this study:

5

a) the emotion strategy: employing emotionally charged terminology in order to sway public opinion, for example, ‘rape’, ‘incest’, ‘murder’ and ‘death’; b) the hypothetical future strategy: explaining ideologies whereby the social actor would describe the consequences in the future of an action taken, or not taken, in the present. It follows the structure: ‘If we don’t do ‘x’ today, ‘y’ will happen tomorrow’; c) the rationality strategy: a decision is presented as having been deeply thought about and the decision-making process said to have been laborious and demanding. It employs phrases such as ‘after a long and hard decision-making process’ and the speaker’s authority is supported by titles, such as, ‘commander’, ‘advisor’, ‘doctor’; d) the voices of expertise strategy: experts in the field in question are employed as a tool in discursive persuasion. When speaking of a medical issue, for example, doctors or consultants are mentioned as supporters. Within this strategy, numbers, percentages, fractions and statistics are important tools to reinforce the legitimization; e) the altruism strategy: ideas and opinions are presented as being of benefit to the community and not spurred on by personal interests. Examples given by Reyes (2011) include democracy, equality and freedom of expression (p. 802).

The visual discourse of the posters is examined according to the framework set out by Ledin and Machin (2018), whereby aspects of the image including setting, colour and objects presented are examined in detail. It is based on the general distinction between ‘denotation’ understood as referring to the basic level of analysis, namely to what is depicted in the image, and ‘connotation’ understood as referring to the meaning of what is presented. This distinction is further broken down into three main points:

a) objects: what objects are represented or not represented in the image; b) colour: how colour is used to “link elements that would otherwise be of different kinds, create bonds or contrasts, evoke moods and associations” (Ledin & Machin, 2018, p. 49). This can also refer to different techniques such as colour choice, saturation to indicate vibrancy or energy and the use of light and shadow. Coordination plays an important role within this category, for example, images may be of the same colour as the logo of an organisation. c) setting: where the image has been taken. This can be used to create contrasts for example by inserting objects not commonly found in the particular setting depicted.

6

Images containing representations of people are analysed with reference to van Leeuwen’s (2008) framework. Aspects of van Leeuwen’s framework employed in this study are: the importance of camera angles and the position of the person in the image in relation to the viewer, giving a sense of involvement (for example, with a close-range picture) in the context of the image or, conversely, detachment. Moreover, the following five factors outlined by van Leeuwen are included in the study:

a) exclusion: this interprets not only what is represented in the image but also what is not there; b) roles: what are the people in the images doing? Do they have a specific role in the image and is this representative of their role in society? This representation could reflect a form of ‘symbolic oppression’ precluding them from one role while binding them to another; c) specific and generic: is the image focusing on one specific person or is it referring to the generic societal group to which the person belongs? d) individuals and groups: people depicted in groups are thereby homogenized and diminished in their individual differences in contrast to images of individuals; e) categorization: visual categorization, for example the depiction of women wearing headscarves or a hijab can be enough to “connote the negative or positive values and associations attached to a particular sociocultural group by the sociocultural group for which the representation is in the first place produced” (van Leeuwen, 2008, p. 145).

The study employed all the strategies of this framework, however for the purpose of this article only selected strategies were incorporated. The following section will present and discuss the results on the basis of the methodological tools presented above.

4. Results and Discussion 4.1 Strategies of Legitimization Employed in Language The analysis of the textual data of both voter groups pre- and post-referendum shows that the preferred strategy in all cases is the use of emotion. This is followed by the category ‘other’ (comments which do not directly adhere to one of the previously mentioned five strategies of legitimization set out by Reyes (2011)) and the remaining strategies have quite a similar distribution between them. Among the ‘No’ voter group (Figure 1), there is little change in the

7 hypothetical future, rationality, voices of expertise and other strategies, but the use of emotion and altruism changes significantly over the period of the study.

Strategies of Legitimization - 'No' Voter Group 60% 54% 50%

40% 33% 28% 30% 22%

Frequency 20% 13% 11%10% 10% 11% Pre-Referendum 10% 5%4% Post-Referendum 1% 0%

Strategies of Legitimization

Figure 1: Strategies of Legitimization employed by ‘No’ voters both before and after the referendum.

The strategies employed by the ‘Yes’ voter group (Figure 2) yield similar results. The preferred strategy is again emotion; however, significant variability is found in the emotion, voices of expertise and altruism strategies.

8

Strategies of Legitimization - 'Yes' Voter Group 60% 51% 50% 41% 40%

30% 22% 18%

20% 15% 14% Frequency 8% 8% 9% 10% 5% 4% 4% Pre-Referendum 0% Post-Referendum

Strategies of Legitimization

Figure 2: Strategies of Legitimization employed by ‘Yes’ voters both before and after the referendum.

a) The emotion strategy

The emotion strategy sees the biggest shift, increasing by 21% after the referendum among the ‘No’ voter group and by 10% among the ‘Yes’ voter group. Prior to the referendum, the ‘No’ voter group uses language which provokes emotions to evoke a reaction and convince others to vote no. These comments consistently utilise specific vocabulary such as ‘kill’, ‘baby’, ‘murder’, ‘shame’. This tendency continues after the referendum, possibly due to a higher emotional sensitivity after having lost to the ‘Yes’ vote.

User 12: Lie on a table and get ready for someone to invade your body, vacuuming out your child, piece by piece, and you’re then left to cramp and bleed and think about what just happened…

The use of the word “invade” in the above comment implies force and the idea that abortion will be performed without the woman’s consent. Child dismemberment is referenced in the phrase “piece by piece” followed by feelings of guilt in the aftermath of the procedure, all intended to provoke a strong emotive reaction in the reader.

The data reveal two sub-strategies within the comments categorised in the emotion strategy:

i. comments attempting to provoke an emotional reaction;

9

ii. comments engaging in the ‘othering’ of another group.

The ‘No’ voter group presents more of a tendency to employ the strategy ‘emotional reaction’ rather than ‘othering’ in the month leading up to the referendum whereas there is a pivot in the month following the referendum, with ‘othering’ becoming the preferred strategy. 21% of the comments belonging to the category emotion before the referendum are engaged in ‘othering’ compared to 58% post-referendum. This could be due to a feeling of resentment towards people who voted ‘Yes’ and towards the people who in the future will turn to abortion as a way of dealing with an unwanted pregnancy.

User 198: Those who vote yes must take responsibility for every baby murdered now.

In addition to this, the distinct groups being othered also change. In the month prior to the referendum, the ‘NO’ voter group being othered is exclusively politicians, given that they have the authority to initiate the referendum process:

User 6: Vote NO and show these devious politicians that we the people still have the power on what goes on in our Country

By contrast, the month following the referendum shows not only politicians being othered but also the ‘Yes’ voters and women.

User 226: It’s an insult to women. I guess they needed attention for there [sic] bodies and choices so anything will do even if it means killing your children and your grandchildren.

Among the ‘Yes’ voter group there is a similar trend as the use of the emotional reaction strategy is significantly higher pre-referendum, for example:

User 146: no child should have to be told mummy might die, when she could have a termination that she needs…no respect for mum…the two boys have suffered enough by losing their dad [in response to an article about a woman diagnosed with heart failure being refused an abortion]

Decreased usage of this strategy suggests the ‘Yes’ voters felt less of a need to evoke an emotional reaction post-referendum.

10

Similar to the ‘No’ voter group, the rate of ‘othering’ increases considerably post-referendum. The target group being othered remains the ‘No’ voter group both before and after the referendum.

User 325: Negative comments from No voters wont [sic] take this smile off my face. Well done Ireland, we made history.

b) The hypothetical future strategy

The hypothetical future strategy remains constant for ‘No’ voters whereas for ‘Yes’ voters there is an increase of 3% post-referendum. This can be interpreted as the consequence of an increased sense of hope for the future of medical care for women in Ireland:

User 334: it won’t happen to any other mother to be brought to court for forced sedation and surgery. The 8th is to be deleted from the HSE’s national “consent” guidelines for all pregnant women!

c) The rationality strategy

Rationality proves an uncommon strategy throughout the campaign for both ‘Yes’ and ‘No’ voters, which leads towards the assumption that voters were rather certain of their voting preferences and did not feel the need to justify their decision. The following is an example of a comment from the ‘No’ voter group post-referendum employing this strategy as the user mentions an extended period of time upon which their judgement has been based:

User 297: I’m a nurse qualified 45yrs [sic] life is precious from birth to death [sic] I’m dual qualified and I have seen it all [sic] don’t dissect the unborn it is vicious [sic] I can’t understand this at the hands of the mother just because it doesn’t suit.

d) The voices of expertise strategy

The strategy voices of expertise (herein VOE) is further subdivided into two categories, VOE positive and VOE negative, referring to authorised speakers in a positive and in a negative sense. The following examples are taken from the ‘No’ voter group comments:

VOE positive:

User 23: Maria was brilliant. A top class woman with great backbone. [Referring to , a barrister and prominent ‘No’ campaigner in Ireland]

VOE negative:

11

User 2: Dr. Death [Referring to Dr. , chairman of the Institute of Gynaecologists and Obstetricians in Ireland]

Negative representation of the authorised speakers among the ‘No’ voter group is prevalent prior to the referendum (67%) whereas after, both categories are more evenly matched (44% positive and 56% negative). This distinction is noteworthy as the same strategy is used but in contrasting ways. The former alludes to authorised speakers with similar views to the user and the latter refers to authorised speakers with opposing views. In both cases, these are utilised in order to justify the standpoint of the user.

In the ‘Yes’ voter group, by contrast, there is a considerably higher percentage of instances referring to authorised speakers with similar views to the commenter before the referendum. This scenario inverts post-referendum with a 37.3% decrease in references to authorised speakers with the same views as the commenter and an identical increase of 37.3% in the usage of reference to authorised speakers with opposing views. This is highlighted in the following comment claiming lack of coherence in the UK Supreme Court’s decision on abortion law in Northern Ireland:

User 309: The [UK] supreme court has dismissed the latest attempt to overturn Northern Ireland’s restrictive abortion laws but nonetheless concluded that they are incompatible with human rights.:O [sic] Literally (straight from the horses' arses): "These laws are incompatible with human rights... but we refuse to overturn them." !!!

e) The altruism strategy

The altruism strategy also witnesses a distinct variation before and after the referendum. Prior to the referendum, 11% of the ‘No’ voters’ comments analysed are in the category altruism, compared to 1% afterwards. This result can be read as an expected hope for the future of the unborn should the Eighth Amendment be retained compared to the disappointment afterwards and the resignation that the unborn will no longer have constitutional protection. Among ‘Yes’ voters’ comments there is an increase of 5% in the use of the altruism strategy post-referendum, which can be interpreted as voters’ expectations for the future of women’s reproductive healthcare after the repeal of the Eighth Amendment:

User 348: So proud of my country for showing true compassion. Thinking tonight of all the previous generations of Irish women and all that they endured and so happy that no future generations will ever have to endure the same.

12

4.2 Discursive Representation of Identities There is a clear difference in the representation of women across the four different data sets. In particular, among the ‘No’ voters post referendum, there are many instances of ‘shaming’ and ‘othering’ of women who chose to obtain an abortion. The underlying opinion of this discourse is that women would choose an abortion for their own convenience without much consideration for the death of the unborn:

User 200: Woman [sic] are so fixated on there [sic] rights. They don’t see that abortion is what we voted for. The murdering and killing of innocent babies.

Abortion in instances of serious cases such as fatal foetal abnormality is not considered by those commenting.

In contrast, before the referendum, similar concerns are presented but women are not directly ‘othered’. Indeed, while in many cases women are being referred to, the words ‘woman’ or ‘female’ are not explicitly mentioned:

User 51: Don’t be fooled by Crocadile [sic] Tears and poor Sob Stories of freedom loving Self Centered animals that cry for the rights to Kill there [sic] own young!

In this dehumanising ‘othering’ of women, the identity that is being imposed on the women in question is that of a self-absorbed, narcissistic, murderous being. This ascribed identity stands in stark opposition to the typical identity associated with the mother figure as maternal, loving, and protective. Moreover, after the referendum there is a tendency to depict the women who had abortions as mentally unstable or “damaged mothers”:

User 189: Where will the psychiatric mothers be placed?

A further interesting finding of the study is the way in which men are portrayed on both sides of the debate. The ‘No’ voters bring attention to the role of the father in the decision to choose abortion over continuing the pregnancy. It is implied that if the Eighth Amendment were repealed women would obtain abortions without including the father in the decision-making process. Men are portrayed as being powerless and choiceless figures while women murder their children:

User 70: Does the father not have a choice of whether his child is murdered or not?

13

Whereas, following the announcement of their decision to vote ‘Yes’ by several prominent male figures in Irish media, the ‘Yes’ voter group ascertains that this alignment with a ‘Yes’ vote constitutes the sign of a ‘real’ man:

User 113: Go the real men of Ireland.

User 118: At least these REAL men have respect for women

This recalls the typical identity associated with men created by a patriarchal system of being strong and protective, and aligns it with advocating for women’s rights, in sharp contrast to the ascribed identity presented by the ‘No’ voter group.

The representation of the unborn varies drastically across both voter groups. The ‘No’ voter group consistently depicts the unborn as innocent ‘babies’, ‘children’ and as being the weakest and most vulnerable members of society, thereby employing the emotional reaction strategy:

User 188: It’s terrible the evil that people can do to beautiful babies.

In stark contrast to this, the ‘Yes’ voter group consistently refers to the unborn in scientific terms using the word ‘foetus’. This indicates a detachment from identifying the foetus that could be aborted as a fully formed baby, i.e. post-birth:

User 166: ruining a young life, terminating a heartbeat, all as if the young foetus was fully formed and conscious of all of its surroundings.

This illustrates the user’s premise of what constitutes a human life, both in terms of physical formation of the body and development of a conscious mind. The following comment criticises the ‘No’ campaign for placing more importance on the life of the foetus than on the life of the mother while claiming to be an advocate for both:

User 136: It’s funny they call it “love both” [name of a ‘No’ campaign] when in fact, it’s more “love the fetus [sic] for better or for worse where the mother’s concerned”.

A further interesting aspect is the diverse use of ‘Ireland’ in its broadest sense in the analysed discourse. Prior to the referendum, the ‘No’ voters when referring to Ireland tend to refer to themselves as being people of Ireland but do not comment on Ireland as a state or on the Eighth Amendment, as the following example reflects:

User 51: Stand up for the Women, Men and Children of Ireland and Vote No!

14

In contrast, after the referendum the ‘No’ voter comments change drastically to consider the removal of the Eighth Amendment to the detriment of Ireland as a nation.

User 241: We good moral respectful proud Irish people will help pick up the damage and destruction that you have caused an Innocent babies [sic] lives from generation to generation.

The image of Ireland that has been created through the discourse of the ‘No’ voters is one of a bleak reality in which babies are murdered for frivolous motivations. This depiction is in stark contrast to that produced in the discourse of the ‘Yes’ side. Before the referendum, ‘Yes’ voters create a similarly negative image of the country of Ireland:

User 104: That our country and Constitution offer no such compassion, no such necessary services to women and families in need, is a stain on our national conscience.

There is also a strong emphasis on Ireland as a state ‘forcing’ citizens to go to Britain to obtain an abortion. This disregards the fact that women also travelled to other countries such as the Netherlands to avail themselves of abortion services (Department of Health & Healthy Ireland, 2015). Given the tumultuous past between Ireland and the United Kingdom, placing the latter in a favourable light with Ireland would be a difficult concept for many Irish citizens to comprehend or accept:

User 110: Thank God that the British looked after our Women when their own Country Cruelly turned their backs on Unfortunate Women with Crisis Pregnancies.

The personification of the nation of Ireland is a recurring trait in the discourse of both ‘Yes’ and ‘No’ voters. The country of Ireland forms the subject of the sentences and therefore, is imagined as a person (personified):

User 324: Goodbye Heartless Ireland! You sent me to London, when my little baby had a fatal foetal abnormality! [underlining by authors]

There are clear identities imposed on the different voter groups in the referendum. The ‘No’ voter group, in the period before the referendum, refer to the ‘Yes’ voters as “turncoats” thereby linking the ‘Yes’ vote to committing treason to the country:

User 5: Shame on them. Vote No to save our future babies from the culture of death that these turncoats are promoting.

15

Both sides of the discussion allude to the opposite voter group as being mentally unstable:

User 100: Too many people live in lala land here. [‘No’ voter group]

User 238: You lot really are away with the fairies aren’t you… [‘Yes’ voter group]

In summary, the identities discursively constructed in the data include those of women, men, Ireland as a nation and the individual voter groups.

4.3 Visual Analysis The visual analysis of twenty-eight campaign posters around Ireland serve to examine the regulated visual discourse employed by both the ‘No’ and the ‘Yes’ campaigning bodies and their role in upholding or refuting the strategies and identities created in the unregulated online discourse. These posters originated from both campaigning bodies and political parties. The ‘No’ posters were created by the campaigning bodies ‘Love Both’ and ‘Save the 8th’ with one poster originating from the political party ‘Aontú’. The ‘Yes’ posters were created by two campaigning bodies ‘ROSA’ and ‘Together for Yes’, and four political parties, ‘Solidarity’, ‘Labour’, ‘Green Party’ and ‘People Before Profit’. These are first analysed according to the framework set out by Ledin and Machin (2018) to examine what is materially represented in the images and the connotations associated with these representations. Secondly, the posters containing images of people are analysed according to the framework devised by van Leeuwen (2008).

The initial key finding of the analysis of the visual discourse is the difference in the presence of people in the images. 73% of the ‘No’ campaign posters contain images of human beings. Posters containing ultrasound images and therefore a likeness to a human baby are included in this classification. The ‘Yes’ campaign, however, includes images of people in only 18% of the posters analysed.

Three of the ‘No’ campaign posters contain stills from ultrasounds. They range from black and white images which would have been more common in past years, to 3D colour images from newer ultrasound techniques. This could be interpreted as a purposeful strategy to appeal to the emotions of both people of older generations who have had children to remind them of what they saw in the ultrasound scans of their children and also to parents of more recent generations. In most cases, it was expected that ultrasound images would have instilled feelings of happiness and anticipation of the impending arrival. Two of these posters contain the technique of

16 personalisation of the unborn with captions such as “My heart started beating at 22 days”, again appealing to the emotions of the viewer.

Interestingly, two of the posters analysed (Figures 3 and 4) share similarities in employing visual techniques while being on opposite sides of the referendum.

Figure 3: Depiction of a woman’s abdomen to advocate a ‘No’ vote.

17

Figure 4: Depiction of a woman’s abdomen to advocate a ‘Yes’ vote.

Both show a woman’s torso, uncovered, with a painted abdominal area. The poster advocating a ‘No’ vote shows a painting on the abdomen of a pregnant woman of a baby in utero with the umbilical cord shaped to form a heart. This is in sharp contrast to the ‘Yes’ vote poster which has the words “My Body” painted on a woman’s uncovered abdomen to form part of the full slogan with the imperative “Stop Policing My Body”. It is important to note she is not visibly pregnant, pointing to the underlying message that the woman in the ‘Yes’ poster had the possibility to decide whether to continue with a pregnancy. In both cases, we do not see the face of the woman depicted, indicating depersonalisation. However, the image is taken in close proximity to the subject, representing intimacy or suggesting a direct involvement of the viewer, which is in strong contrast to posters with missing faces that instead convey the idea that the woman could be anyone, including the viewer.

In total, three of the ‘Yes’ campaign posters show images of people. It needs to be highlighted that none of the ‘Yes’ posters contains imagery of babies or children. The second of these posters shows a woman’s face with a concerned expression looking into the distance. She is not looking directly at the viewer and so there is a sense of detachment. This could be

18 interpreted as looking to the future in the hopes of creating new legislation in Ireland. The caption “Your sister, your friend, your daughter” gives the idea that the woman in the image could be someone related to the viewer, thereby evoking an emotional response.

The third image of a person on the ‘Yes’ campaign posters depicts Savita Halappanavar. She was admitted to hospital in 2012 with the diagnosis of an impending miscarriage and was refused an abortion under the Eighth Amendment as there was still the presence of a foetal heartbeat. She died as a result of undiagnosed sepsis.

The use of Savita’s image in the ‘Yes’ campaign is extremely significant as her death brought the restrictions of the Eighth Amendment to light not only in Ireland but in the world. This led to the intervention of the United Nations to ensure Ireland revisit the existing legislation and ultimately, propose this referendum to repeal the Eighth Amendment in the Constitution.

Figure 5: Image of four Republican leaders and a reference to the Proclamation used to advocate a ‘No’ vote.

One further example of the ‘No’ campaign posters (Figure 5) depicts four Republican leaders alongside the caption “They said to cherish all the children equally” (underlining on poster). This is a quote from the Proclamation of Independence, a document published in 1916 denoting the beginning of the Easter Rising in Ireland and ultimately, emancipation from British rule to create the Republic of Ireland. The image also contains a backdrop of the Irish flag. These

19 components have the power to instil not only a sense of national pride in the viewer but also a sense of rebellion (see also ‘join the rebellion’ in Figure 3). This poster could also link to the voices of expertise strategy of legitimization as outlined by Reyes (2011).

The contrasting of Britain and Ireland occurs on a few occasions. Although Britain is not explicitly mentioned in the previous poster, this implication would be understood by Irish viewers. In other posters there is an explicit positioning of Ireland versus Britain. Captions reflecting this include: “Join the Rebellion”, “In England 1 in 5 babies are aborted. Don’t bring this to Ireland”. All of these would be categorised under the emotion strategy of legitimization as for many Irish people there are strong emotions associated with the independence of the Republic of Ireland and the battles fought to obtain it.

The second key finding of the visual analysis is the usage of colour in the posters. The colour white is the most utilized colour appearing on 100% of the ‘No’ posters and 88% of the ‘Yes’ posters. Studies into the emotional impact of the colour white have determined it to be associated with positive connotations due to its “lightness”, transmitting a sense of tranquillity, cleanliness, innocence and purity (Gil & Le Bigot, 2014, p. 11). Kauppinen-Räisänen and Jauffret (2018), in contrast, mention that in some countries the colour white can be associated with mourning or horror, given the expression “white as a sheet” after a traumatic experience (p. 111). However one could interpret it also as a pragmatic choice. Thus, it proves to be difficult to pinpoint the exact motivations underlying this choice by both sides of the campaign.

In general, the discourse presented in white is of a shocking nature, as exemplified by the captions “170,000+ forced abroad for abortions” (‘Yes’ campaign) and “In Britain, 90% of babies with Down Syndrome are aborted” (‘No’ campaign). Therefore, it could be suggested that the choice of colour was made not only to inform but to provoke strong emotional reactions in people with either what was happening in Ireland at the time of the campaign (for the ‘Yes’ side) and what would happen in the future if the Eighth Amendment were to be repealed (for the ‘No’ side). It can hence be concluded that the colour white is employed with the intention of evoking emotions in the viewer.

Remarkably, the colour purple proves to be popular for the ‘Yes’ campaign as it is featured in 53% of the posters analysed. While the colour purple is often associated with contemporary feminism, in antiquity it was commonly used to convey royalty and prestige (Elliott, 2008, p. 176). It is noteworthy that all posters with purple features put women in their focus. This includes an anecdote by women affected by the Eighth Amendment, representations of women

20 or explicit exclamations regarding women, for example, in the phrase “Trust Women”. This could lead to the interpretation that in encouraging voters to trust women presented by the colour purple that women are in a position of prestige or will be in a position of prestige if the Eighth Amendment is repealed. While the ‘No’ campaign does not use the colour purple in any of their posters, they rely heavily on the colour pink. Pink can be used to convey positive information and is often associated with femininity (Gil & Le Bigot, 2014, p. 9). Koller (2008) highlights that pink can be “connected with femininity and its stereotypical features, such as softness and delicacy, with childhood and innocence as well as with vanity and artificiality”, however “there are also emergent and less obvious connotations like independence and fun, sexuality and lust” (p. 396).

The visual analysis clearly reveals that the data analysed is mainly supportive of the findings derived from the textual data. Firstly, the dehumanization of the unborn using words such as “foetus” and the lack of people and babies depicted in the ‘Yes’ posters both suggest an emotional detachment from the issue of abortion. This is in direct contrast to the use of the words “baby” and “child” to describe the unborn; and the high frequency of images of people incorporated in the ‘No’ posters. Secondly, the strategies employed in the posters in an attempt to persuade voters are also reflective of the textual discourse analysis in that the most common strategy is the strategy of emotion. The depiction of the unborn as a baby, already with human characteristics, strives to heighten an emotional reaction in the viewer. Finally, the lack of the colour purple in the ‘No’ campaign posters, given its connotations of prestige and feminism, coincides with the tendency in the textual data towards ‘othering’ and ‘shaming’ women who choose to have an abortion. Purple, in contrast, is frequently utilized by the ‘Yes’ campaign in posters praising women, which reflects the textual discourse requesting that women are no longer to be placed at equal importance with the unborn.

5. Conclusion This article sheds light on regulated and unregulated discourse of the campaign surrounding the referendum to Repeal the Eighth Amendment to the Constitution of Ireland in 2018. On the basis of Reyes’ (2011) strategies of legitimization, the key finding of the textual analysis is that the emotional strategy is the most prominent category in use, while the strategies hypothetical future, rationality, voices of expertise and altruism are used significantly less in both ‘Yes’ and ‘No’ voters’ groups prior to and after the referendum. It reveals the high emotional involvement

21 of voters of each side which outnumbers all the other possible strategies. It needs to be emphasized that a considerable number of discursive strategies could not be linked to any of Reyes’ (2011) strategies which clearly reveals that there is a need for further categorisation and/or subcategorization. Some of the subcategories for strategies are introduced here, for instance, the study reveals that the category emotion needs to be subdivided into strategies that provoke an emotional reaction and into those which express othering. Concerning the construction of identities, the discourse reveals that both the construction of positive and negative ‘mother identities’ and ‘voter identities’ as well as the reference to the national identity are of high significance.

The visual analysis supports the textual analysis insofar as it also employs the strategy emotion to a high extent, often done via the depiction of various ultrasounds. Moreover, it becomes clear that the visual discourse utilizes colour purposefully in order to support a specific discourse.

To conclude, the purpose of the study was to analyse multimodal discursive strategies employed around the referendum to Repeal the Eighth Amendment in Ireland in 2018. In order to be able to draw general conclusions, a more comprehensive examination of discursive strategies employed around the topic of abortion in Ireland is needed. Additionally, it is suggested that a comparative cross-cultural study would lead to further interesting insights.

References Carnegie, A., & Roth, R. (2019). From the grassroots to the Oireachtas: Abortion law reform in the Republic of Ireland. Health and Human Rights, 21(2), 109-120. Creswell, J. W., & Plano Clark, V. L. (2007). Designing and conducting mixed methods research. Sage Publications. Department of Health & Healthy Ireland. (2015). National Sexual Health Strategy 2015 - 2020. https://assets.gov.ie/7562/e5a5ac26eb22405aaf6538656564690a.pdf Elkink, J. A., Farrell, D. M., Marien, S., Reidy, T., & Suiter, J. (2020). The death of conservative Ireland? The 2018 abortion referendum. Electoral Studies, 65, 1-11. https://doi.org/10.1016/j.electstud.2020.102142 Elliott, C. (2008). Purple pasts: Color codification in the ancient world. Law & Social Inquiry, 33(1), 173–194. https://doi.org/10.1111/j.1747-4469.2008.00097.x Fairclough, N. (1989). Language and power. Routledge.

22

Field, L. (2018). The abortion referendum of 2018 and a timeline of abortion politics in Ireland to date. Irish Political Studies, 33(4), 608-628. https://doi.org/full/10.1080/07907184.2018.1500461 Gil, S., & Le Bigot, L. (2014). Seeing life through positive-tinted glasses: Color-meaning associations. PLoS ONE, 9(8). https://doi.org/10.1371/journal.pone.0104291 Jäger, S. (2001). Discourse and knowledge: Theoretical and methodological aspects of a critical discourse and dispositive analysis. In R. Wodak & M. Meyer (Eds.), Methods of critical discourse analysis (pp. 32-62). Sage Publications. Kauppinen-Räisänen, H., & Jauffret, M. N. (2018). Using colour semiotics to explore colour meanings. Qualitative Market Research, 21(1), 101-117. https://doi.org/10.1108/QMR- 03-2016-0033 Koller, V. (2008). ‘Not just a color’: Pink as a gender and sexuality marker in visual communication. Visual Communication. 7(4), 395-423. https://doi.org/10.1177/1470357208096209 Ledin, P., & Machin, D. (2018). Doing visual analysis. Sage Publications. Reyes, A. (2011). Strategies of legitimization in political discourse: From words to actions. Discourse & Society, 22(6), 781–807. https://doi.org/10.1177/0957926511419927 Smyth, L. (2005). Abortion and nation: The politics of reproduction in contemporary Ireland. Ashgate Publishing Limited. Van Leeuwen, T. (2008). Discourse and practice: New tools for critical discourse analysis. Oxford University Press. https://doi.org/10.1093/acprof Wodak, R. (2011). “Us” and “them”: Inclusion/exclusion – Discrimination via discourse. In G. Delanty, R. Wodak, & P. Jones (Eds.), Migration, identity and belonging (pp. 54–77). Liverpool University Press.